The venerable Western Hemlock - Washington’s state tree - has been in decline in the Puget Lowlands since 2015. The trees are dying - not everywhere at at once, but intermittently, often in clusters, throughout low elevations west of the Cascades.

Researchers initially attributed the mortality to Rhizoctonia butini, a fungus in the order Cantharellales. This hypothesis has since fallen out of favor (Jared LeBoldus, OSU Corvallis, personal communication). No subsequent explanation has emerged.

We frequently hear this mortality attributed to normal old age senescence, or climate change.

We are not so sure. Casual observation in Seward Park’s 120-acre old-growth urban forest suggested that age and climate did not explain hemlock mortality, in that we see areas with many dead trees, and we see many more areas dense with healthy trees. A mix of tree ages is found in both kinds of sites. We can assume that climate and weather patterns are constant across this small forest.

If the observed contrast holds up, and if the population of trees in the two contrasting areas are otherwise identical - by age, density, and site characteristics - then we have an ideal site for studying the regional decline. Seward provides a natural experiment.

Students Mapped, Measured and Assessed 724 Hemlock Trees

Between January and March, two Univeristy of Washington students from the Program on the Environment in their senior Capstone Research Project, along with a Garfield High School junior spent about nine weeks in the forest, collecting data to test our hypothesis.

We also used data collected in the summer of 2021 by CHOOSE 180 interns.

Our citizen and student science sets the stage for in-depth research. These are the steps we have taken:

  1. We initially identified a two acre healthy site (the “hemlock garden”) to compare to the six acre affected site (the “hemlock graveyard”)
  2. We mapped, measured and assessed the health of all trees in these two sites
  3. We found that these two sites, separated within continuous forest by 700 meters, were identical except for mortality: same size distribution, weather patterns, and soil type.
  4. We next expanded our survey to include 724 total hemlocks across most of the 120-acre forest
  5. Analysis (significance tests and kernel density calculations, shown below) establish that the hemlock graveyard is a significant outlier with respect to hemlock mortality in this forest.

Initial Evidence From Small Sites: “Graveyard” and “Garden”

We started, as mentioned above, with all the trees in two dense hemlock sites. Preliminary statistics confirmed our intuition, as these boxplots show.

Subsequent mapping and analysis refined these results - see below.

We soon realized the importance of broadening our study. Here is a screenshot of our efforts at an intermediate stage, the blue circles marking the garden and graveyard, with many additional trees added.

In subsequent work, shown below, the blue, visually estimated boundaries are replaced by analytical boundaries calculated by kernel density methods.

Density Maps: All Hemlocks; dying and dead; healthh hemlocks

“Overplotting” - the display of abundant overlapping points - makes it hard (quite literally), to see the forest for the trees. We see this in the map above, and in the final interactive version available here. Kernel density mapping solves the problem: large patterns to emerge. These three panels below do this, showing the concentration of all (left), of affected (middle) and of healthy (left) hemlocks across the forest.

Selecting Contrasting Candidate Sites for Further Study

These three density maps confirm our informal observation, and hypothesis, that contrasting health and Seward’s version of hemlock decline is not straightforwardly explained by senesence or weather patterns.

To strengthen our case, we defined two contiguous regions with

  1. similar density of hemlocks
  2. contrasting health status
Our method:
  1. subset the full observation table by health
  2. calculate kernel densities of hemlocks in both groups
  3. from these densities, identify the boundaries of comparable sites 
## [1] NA NA

Density-derived Boundaries for Hemlock Graveyard and Garden

Here you can see that healthy hemlocks are dense thoughout the central and southern regions of the forest, in an area that includes, but extends far beyong our original “hemlock garden” site.

Conclusion

Our citizen and student science establishes that in Seward Park’s old-growth forest, hemlock die-off is dramatic; is not caused simply by senescence or weather patterns, and that it provides an opportunity for further (professional) study.

We predict that foliar and mycorrhizal DNA sequencing, entomological studies, and root rot analyses may all be needed to understand this problem. Once understood, remedies may become possible.